I completed this lab on 9/13/22 in order to act as a useful reference for future practice.

My implementation included some slight differences compared to what you might normally implement for the lab, so I wanted to detail them here.

The lab notes here that it is recommended to store the terraform state for the infrastructure in an S3 bucket, although it is not necessary. HOWEVER, that only applies to the infra repository! For this repository, having the terraform state managed in an S3 bucket is strictly required. The reason for this is that we intend the pipeline itself to execute terraform commands on our behalf. This means the pipeline needs access to the terraform state to perform modifications.

Since AWS CodePipeline build runners are VMs that shut down after the pipeline has completed, that means any local terraform state would be lost. The solution to this is to store your terraform state in an S3 bucket, and update the configuration accordingly. This means that it is generally recommended to go back to Step 1, and incorporate shared state for the entire lab.

It is important to note however, that instead of creating an entire S3 bucket configured from scratch to do so, I used an S3 bucket that pre-existed. I have previously created an S3 bucket called "revature-terraform" for storing my terraform state for other projects. That bucket followed all of the general best-practices regarding securing your state, such as bucket versioning, encryption, locking the statefile, etc.

While completing your lab, you won't have access to a pre-existing bucket to store your state in, so you will have to construct it yourself. Terraform has documentation regarding this here. Additionally, a fairly recent video tutorial regarding setting this up can be found here. NOTE: I haven't watched the video in detail, so I make no promises to exactly how useful it is, but I have made a cursory glance-through and it seems alright.

I have opened a Pull Request on the repository that includes every single change I made to the infrastructure, with comments explaining what I was doing and why. This can be found and reviewed here.